GitHub/LineageOS/G12/android_kernel_amlogic_linux-4.9.git
13 years agomlx4: Add blue flame support for kernel consumers
Eli Cohen [Tue, 22 Mar 2011 22:38:41 +0000 (22:38 +0000)]
mlx4: Add blue flame support for kernel consumers

Using blue flame can improve latency by allowing the HW to more efficiently
access the WQE. This patch presents two functions that are used to allocate or
release HW resources for using blue flame; the caller need to supply a struct
mlx4_bf object when allocating resources. Consumers that make use of this API
should post doorbells to the UAR object pointed by the initialized struct
mlx4_bf;

Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: Enabling new steering
Yevgeny Petrilin [Tue, 22 Mar 2011 22:38:31 +0000 (22:38 +0000)]
mlx4_en: Enabling new steering

The mlx4_en module now uses the new steering mechanism.
The RX packets are now steered through the MCG table instead
of Mac table for unicast, and default entry for multicast.
The feature is enabled through INIT_HCA

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4: Add support for promiscuous mode in the new steering model.
Yevgeny Petrilin [Tue, 22 Mar 2011 22:38:24 +0000 (22:38 +0000)]
mlx4: Add support for promiscuous mode in the new steering model.

For Ethernet mode only,
When we want to register QP as promiscuous, it must be added to all the
existing steering entries and also to the default one.
The promiscuous QP might also be on of "real" QPs,
which means we need to monitor every entry to avoid duplicates and ensure
we close an entry when all it has is promiscuous QPs.
Same mechanism both for unicast and multicast.

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4: generalization of multicast steering.
Yevgeny Petrilin [Tue, 22 Mar 2011 22:38:17 +0000 (22:38 +0000)]
mlx4: generalization of multicast steering.

The same packet steering mechanism would be used both for IB and Ethernet,
Both multicasts and unicasts.
This commit prepares the general infrastructure for this.

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: Reporting HW revision in ethtool -i
Yevgeny Petrilin [Tue, 22 Mar 2011 22:38:07 +0000 (22:38 +0000)]
mlx4_en: Reporting HW revision in ethtool -i

HW revision is derived from device ID and rev id.

Signed-off-by: Eugenia Emantayev <eugenia@mellanox.co.il>
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4: Wake on LAN support
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:59 +0000 (22:37 +0000)]
mlx4: Wake on LAN support

The driver queries the FW for WOL support.
Ethtool get/set_wol is implemented accordingly.
Only magic packets are supported at the time.

Signed-off-by: Igor Yarovinsky <igory@mellanox.co.il>
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: using new mlx4 interrupt scheme
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:52 +0000 (22:37 +0000)]
mlx4_en: using new mlx4 interrupt scheme

Each RX ring will have its own interrupt vector, and TX rings will share one
(we mostly use polling for TX completions).
The vectors are assigned first time device is opened, and its name includes
the interface name and ring number.

Signed-off-by: Markuze Alex <markuze@mellanox.co.il>
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4: Changing interrupt scheme
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:47 +0000 (22:37 +0000)]
mlx4: Changing interrupt scheme

Adding a pool of MSI-X vectors and EQs that can be used explicitly by mlx4_core
customers (mlx4_ib, mlx4_en). The consumers will assign their own names to the
interrupt vectors. Those vectors are not opened at mlx4 device initialization,
opened by demand.
Changed the max number of possible EQs according to the new scheme, no longer relies on
on number of cores.
The new functionality is exposed through mlx4_assign_eq() and mlx4_release_eq().
Customers that do not use the new API will get completion vectors as before.

Signed-off-by: Markuze Alex <markuze@mellanox.co.il>
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: bringing link up when registering netdevice
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:41 +0000 (22:37 +0000)]
mlx4_en: bringing link up when registering netdevice

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: optimize adaptive moderation algorithm for better latency
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:36 +0000 (22:37 +0000)]
mlx4_en: optimize adaptive moderation algorithm for better latency

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: moderation parameters are not reseted.
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:28 +0000 (22:37 +0000)]
mlx4_en: moderation parameters are not reseted.

Instead of reseting the module parameters each ifup or mtu change,
they are being set once at device initialization
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomlx4_en: going out of range of TX rings when reporting stats
Yevgeny Petrilin [Tue, 22 Mar 2011 22:37:23 +0000 (22:37 +0000)]
mlx4_en: going out of range of TX rings when reporting stats

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipv4: fix ip_rt_update_pmtu()
Eric Dumazet [Wed, 23 Mar 2011 19:18:15 +0000 (12:18 -0700)]
ipv4: fix ip_rt_update_pmtu()

commit 2c8cec5c10bc (Cache learned PMTU information in inetpeer) added
an extra inet_putpeer() call in ip_rt_update_pmtu().

This results in various problems, since we can free one inetpeer, while
it is still in use.

Ref: http://www.spinics.net/lists/netdev/msg159121.html

Reported-by: Alexander Beregalov <a.beregalov@gmail.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipv4: Fallback to FIB local table in __ip_dev_find().
David S. Miller [Wed, 23 Mar 2011 04:56:23 +0000 (21:56 -0700)]
ipv4: Fallback to FIB local table in __ip_dev_find().

In commit 9435eb1cf0b76b323019cebf8d16762a50a12a19
("ipv4: Implement __ip_dev_find using new interface address hash.")
we reimplemented __ip_dev_find() so that it doesn't have to
do a full FIB table lookup.

Instead, it consults a hash table of addresses configured to
interfaces.

This works identically to the old code in all except one case,
and that is for loopback subnets.

The old code would match the loopback device for any IP address
that falls within a subnet configured to the loopback device.

Handle this corner case by doing the FIB lookup.

We could implement this via inet_addr_onlink() but:

1) Someone could configure many addresses to loopback and
   inet_addr_onlink() is a simple list traversal.

2) We know the old code works.

Reported-by: Julian Anastasov <ja@ssi.bg>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: Make undo_ssthresh arg to tcp_undo_cwr() a bool.
David S. Miller [Wed, 23 Mar 2011 02:37:11 +0000 (19:37 -0700)]
tcp: Make undo_ssthresh arg to tcp_undo_cwr() a bool.

Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agotcp: avoid cwnd moderation in undo
Yuchung Cheng [Mon, 14 Mar 2011 10:57:03 +0000 (10:57 +0000)]
tcp: avoid cwnd moderation in undo

In the current undo logic, cwnd is moderated after it was restored
to the value prior entering fast-recovery. It was moderated first
in tcp_try_undo_recovery then again in tcp_complete_cwr.

Since the undo indicates recovery was false, these moderations
are not necessary. If the undo is triggered when most of the
outstanding data have been acknowledged, the (restored) cwnd is
falsely pulled down to a small value.

This patch removes these cwnd moderations if cwnd is undone
  a) during fast-recovery
b) by receiving DSACKs past fast-recovery

Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agobridge: Fix possibly wrong MLD queries' ethernet source address
Linus Lüssing [Tue, 22 Mar 2011 11:40:32 +0000 (11:40 +0000)]
bridge: Fix possibly wrong MLD queries' ethernet source address

The ipv6_dev_get_saddr() is currently called with an uninitialized
destination address. Although in tests it usually seemed to nevertheless
always fetch the right source address, there seems to be a possible race
condition.

Therefore this commit changes this, first setting the destination
address and only after that fetching the source address.

Reported-by: Jan Beulich <JBeulich@novell.com>
Signed-off-by: Linus Lüssing <linus.luessing@web.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: davinci_emac:Fix translation logic for buffer descriptor
Sriram [Tue, 22 Mar 2011 02:31:03 +0000 (02:31 +0000)]
net: davinci_emac:Fix translation logic for buffer descriptor

With recent changes to the driver(switch to new cpdma layer),
the support for buffer descriptor address translation logic
is broken. This affects platforms where the physical address of
the descriptors as seen by the DMA engine is different from the
physical address.

Original Patch adding translation logic support:
Commit: ad021ae8862209864dc8ebd3b7d3a55ce84b9ea2

Signed-off-by: Sriramakrishnan A G <srk@ti.com>
Tested-By: Sekhar Nori <nsekhar@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipv6: ip6_route_output does not modify sk parameter, so make it const
Florian Westphal [Wed, 23 Mar 2011 02:17:36 +0000 (19:17 -0700)]
ipv6: ip6_route_output does not modify sk parameter, so make it const

This avoids explicit cast to avoid 'discards qualifiers'
compiler warning in a netfilter patch that i've been working on.

Signed-off-by: Florian Westphal <fw@strlen.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wirel...
David S. Miller [Tue, 22 Mar 2011 21:36:18 +0000 (14:36 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-2.6

13 years agoipv4: optimize route adding on secondary promotion
Julian Anastasov [Sat, 19 Mar 2011 12:13:54 +0000 (12:13 +0000)]
ipv4: optimize route adding on secondary promotion

Optimize the calling of fib_add_ifaddr for all
secondary addresses after the promoted one to start from
their place, not from the new place of the promoted
secondary. It will save some CPU cycles because we
are sure the promoted secondary was first for the subnet
and all next secondaries do not change their place.

Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipv4: remove the routes on secondary promotion
Julian Anastasov [Sat, 19 Mar 2011 12:13:52 +0000 (12:13 +0000)]
ipv4: remove the routes on secondary promotion

The secondary address promotion relies on fib_sync_down_addr
to remove all routes created for the secondary addresses when
the old primary address is deleted. It does not happen for cases
when the primary address is also in another subnet. Fix that
by deleting local and broadcast routes for all secondaries while
they are on device list and by faking that all addresses from
this subnet are to be deleted. It relies on fib_del_ifaddr being
able to ignore the IPs from the concerned subnet while checking
for duplication.

Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipv4: fix route deletion for IPs on many subnets
Julian Anastasov [Sat, 19 Mar 2011 12:13:49 +0000 (12:13 +0000)]
ipv4: fix route deletion for IPs on many subnets

Alex Sidorenko reported for problems with local
routes left after IP addresses are deleted. It happens
when same IPs are used in more than one subnet for the
device.

Fix fib_del_ifaddr to restrict the checks for duplicate
local and broadcast addresses only to the IFAs that use
our primary IFA or another primary IFA with same address.
And we expect the prefsrc to be matched when the routes
are deleted because it is possible they to differ only by
prefsrc. This patch prevents local and broadcast routes
to be leaked until their primary IP is deleted finally
from the box.

As the secondary address promotion needs to delete
the routes for all secondaries that used the old primary IFA,
add option to ignore these secondaries from the checks and
to assume they are already deleted, so that we can safely
delete the route while these IFAs are still on the device list.

Reported-by: Alex Sidorenko <alexandre.sidorenko@hp.com>
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipv4: match prefsrc when deleting routes
Julian Anastasov [Sat, 19 Mar 2011 12:13:46 +0000 (12:13 +0000)]
ipv4: match prefsrc when deleting routes

fib_table_delete forgets to match the routes by prefsrc.
Callers can specify known IP in fc_prefsrc and we should remove
the exact route. This is needed for cases when same local or
broadcast addresses are used in different subnets and the
routes differ only in prefsrc. All callers that do not provide
fc_prefsrc will ignore the route prefsrc as before and will
delete the first occurence. That is how the ip route del default
magic works.

Current callers are:

- ip_rt_ioctl where rtentry_to_fib_config provides fc_prefsrc only
when the provided device name matches IP label with colon.

- inet_rtm_delroute where RTA_PREFSRC is optional too

- fib_magic which deals with routes when deleting addresses
and where the fc_prefsrc is always set with the primary IP
for the concerned IFA.

Signed-off-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoNET: smsc95xx: don't use stack for async writes to the device
Marc Zyngier [Fri, 18 Mar 2011 03:53:58 +0000 (03:53 +0000)]
NET: smsc95xx: don't use stack for async writes to the device

The set_multicast operation performs asynchronous writes to the
device, with some addresses pointing to the stack. Bad things may
happen, and this is trapped CONFIG_DMA_API_DEBUG:

[    5.237762] WARNING: at /build/buildd/linux-linaro-omap-2.6.38/lib/dma-debug.c:867 check_for_stack+0xd4/0x100()
[    5.237792] ehci-omap ehci-omap.0: DMA-API: device driver maps memory fromstack [addr=d9c77dec]
[    5.237792] Modules linked in: smsc95xx(+) usbnet twl6030_usb twl4030_pwrbutton leds_gpio omap_wdt omap2_mcspi
[    5.237854] [<c006d618>] (unwind_backtrace+0x0/0xf8) from [<c00a6a14>] (warn_slowpath_common+0x54/0x64)
[    5.237884] [<c00a6a14>] (warn_slowpath_common+0x54/0x64) from [<c00a6ab8>] (warn_slowpath_fmt+0x30/0x40)
[    5.237915] [<c00a6ab8>] (warn_slowpath_fmt+0x30/0x40) from [<c034e9d8>] (check_for_stack+0xd4/0x100)
[    5.237915] [<c034e9d8>] (check_for_stack+0xd4/0x100) from [<c034fea8>] (debug_dma_map_page+0xb4/0xdc)
[    5.237976] [<c034fea8>] (debug_dma_map_page+0xb4/0xdc) from [<c04242f0>] (map_urb_for_dma+0x26c/0x304)
[    5.237976] [<c04242f0>] (map_urb_for_dma+0x26c/0x304) from [<c0424594>] (usb_hcd_submit_urb+0x78/0x19c)
[    5.238037] [<c0424594>] (usb_hcd_submit_urb+0x78/0x19c) from [<bf049c5c>] (smsc95xx_write_reg_async+0xb4/0x130 [smsc95xx])
[    5.238067] [<bf049c5c>] (smsc95xx_write_reg_async+0xb4/0x130 [smsc95xx]) from [<bf049dd4>] (smsc95xx_set_multicast+0xfc/0x148 [smsc95xx])
[    5.238098] [<bf049dd4>] (smsc95xx_set_multicast+0xfc/0x148 [smsc95xx]) from [<bf04a118>] (smsc95xx_reset+0x2f8/0x68c [smsc95xx])
[    5.238128] [<bf04a118>] (smsc95xx_reset+0x2f8/0x68c [smsc95xx]) from [<bf04a8cc>] (smsc95xx_bind+0xcc/0x188 [smsc95xx])
[    5.238159] [<bf04a8cc>] (smsc95xx_bind+0xcc/0x188 [smsc95xx]) from [<bf03ef1c>] (usbnet_probe+0x204/0x4c4 [usbnet])
[    5.238220] [<bf03ef1c>] (usbnet_probe+0x204/0x4c4 [usbnet]) from [<c0429078>] (usb_probe_interface+0xe4/0x1c4)
[    5.238250] [<c0429078>] (usb_probe_interface+0xe4/0x1c4) from [<c03a8770>] (really_probe+0x64/0x160)
[    5.238250] [<c03a8770>] (really_probe+0x64/0x160) from [<c03a8a30>] (driver_probe_device+0x48/0x60)
[    5.238281] [<c03a8a30>] (driver_probe_device+0x48/0x60) from [<c03a8ad4>] (__driver_attach+0x8c/0x90)
[    5.238311] [<c03a8ad4>] (__driver_attach+0x8c/0x90) from [<c03a7b24>] (bus_for_each_dev+0x50/0x7c)
[    5.238311] [<c03a7b24>] (bus_for_each_dev+0x50/0x7c) from [<c03a82ec>] (bus_add_driver+0x190/0x250)
[    5.238311] [<c03a82ec>] (bus_add_driver+0x190/0x250) from [<c03a8cf8>] (driver_register+0x78/0x13c)
[    5.238433] [<c03a8cf8>] (driver_register+0x78/0x13c) from [<c0428040>] (usb_register_driver+0x78/0x13c)
[    5.238464] [<c0428040>] (usb_register_driver+0x78/0x13c) from [<c005b680>] (do_one_initcall+0x34/0x188)
[    5.238494] [<c005b680>] (do_one_initcall+0x34/0x188) from [<c00e11f0>] (sys_init_module+0xb0/0x1c0)
[    5.238525] [<c00e11f0>] (sys_init_module+0xb0/0x1c0) from [<c0065c40>] (ret_fast_syscall+0x0/0x30)

Move the two offenders to the private structure which is kmalloc-ed,
and thus safe.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Cc: Steve Glendinning <steve.glendinning@smsc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: implement dev_disable_lro() hw_features compatibility
Michał Mirosław [Fri, 18 Mar 2011 16:56:34 +0000 (16:56 +0000)]
net: implement dev_disable_lro() hw_features compatibility

Implement compatibility with new hw_features for dev_disable_lro().
This is a transition path - dev_disable_lro() should be later
integrated into netdev_fix_features() after all drivers are converted.

Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoIPVS: Use global mutex in ip_vs_app.c
Simon Horman [Mon, 21 Mar 2011 15:18:01 +0000 (15:18 +0000)]
IPVS: Use global mutex in ip_vs_app.c

As part of the work to make IPVS network namespace aware
__ip_vs_app_mutex was replaced by a per-namespace lock,
ipvs->app_mutex. ipvs->app_key is also supplied for debugging purposes.

Unfortunately this implementation results in ipvs->app_key residing
in non-static storage which at the very least causes a lockdep warning.

This patch takes the rather heavy-handed approach of reinstating
__ip_vs_app_mutex which will cover access to the ipvs->list_head
of all network namespaces.

[   12.610000] IPVS: Creating netns size=2456 id=0
[   12.630000] IPVS: Registered protocols (TCP, UDP, SCTP, AH, ESP)
[   12.640000] BUG: key ffff880003bbf1a0 not in .data!
[   12.640000] ------------[ cut here ]------------
[   12.640000] WARNING: at kernel/lockdep.c:2701 lockdep_init_map+0x37b/0x570()
[   12.640000] Hardware name: Bochs
[   12.640000] Pid: 1, comm: swapper Tainted: G        W 2.6.38-kexec-06330-g69b7efe-dirty #122
[   12.650000] Call Trace:
[   12.650000]  [<ffffffff8102e685>] warn_slowpath_common+0x75/0xb0
[   12.650000]  [<ffffffff8102e6d5>] warn_slowpath_null+0x15/0x20
[   12.650000]  [<ffffffff8105967b>] lockdep_init_map+0x37b/0x570
[   12.650000]  [<ffffffff8105829d>] ? trace_hardirqs_on+0xd/0x10
[   12.650000]  [<ffffffff81055ad8>] debug_mutex_init+0x38/0x50
[   12.650000]  [<ffffffff8104bc4c>] __mutex_init+0x5c/0x70
[   12.650000]  [<ffffffff81685ee7>] __ip_vs_app_init+0x64/0x86
[   12.660000]  [<ffffffff81685a3b>] ? ip_vs_init+0x0/0xff
[   12.660000]  [<ffffffff811b1c33>] T.620+0x43/0x170
[   12.660000]  [<ffffffff811b1e9a>] ? register_pernet_subsys+0x1a/0x40
[   12.660000]  [<ffffffff81685a3b>] ? ip_vs_init+0x0/0xff
[   12.660000]  [<ffffffff81685a3b>] ? ip_vs_init+0x0/0xff
[   12.660000]  [<ffffffff811b1db7>] register_pernet_operations+0x57/0xb0
[   12.660000]  [<ffffffff81685a3b>] ? ip_vs_init+0x0/0xff
[   12.670000]  [<ffffffff811b1ea9>] register_pernet_subsys+0x29/0x40
[   12.670000]  [<ffffffff81685f19>] ip_vs_app_init+0x10/0x12
[   12.670000]  [<ffffffff81685a87>] ip_vs_init+0x4c/0xff
[   12.670000]  [<ffffffff8166562c>] do_one_initcall+0x7a/0x12e
[   12.670000]  [<ffffffff8166583e>] kernel_init+0x13e/0x1c2
[   12.670000]  [<ffffffff8128c134>] kernel_thread_helper+0x4/0x10
[   12.670000]  [<ffffffff8128ad40>] ? restore_args+0x0/0x30
[   12.680000]  [<ffffffff81665700>] ? kernel_init+0x0/0x1c2
[   12.680000]  [<ffffffff8128c130>] ? kernel_thread_helper+0x0/0x1global0

Signed-off-by: Simon Horman <horms@verge.net.au>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Julian Anastasov <ja@ssi.bg>
Cc: Hans Schillstrom <hans@schillstrom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipvs: fix a typo in __ip_vs_control_init()
Eric Dumazet [Mon, 21 Mar 2011 10:15:40 +0000 (10:15 +0000)]
ipvs: fix a typo in __ip_vs_control_init()

Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Simon Horman <horms@verge.net.au>
Cc: Julian Anastasov <ja@ssi.bg>
Acked-by: Simon Horman <horms@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoveth: Fix the byte counters
Eric W. Biederman [Tue, 22 Mar 2011 01:24:53 +0000 (18:24 -0700)]
veth: Fix the byte counters

Commit 44540960 "veth: move loopback logic to common location" introduced
a bug in the packet counters.  I don't understand why that happened as it
is not explained in the comments and the mut check in dev_forward_skb
retains the assumption that skb->len is the total length of the packet.

I just measured this emperically by setting up a veth pair between two
noop network namespaces setting and attempting a telnet connection between
the two.  I saw three packets in each direction and the byte counters were
exactly 14*3 = 42 bytes high in each direction.  I got the actual
packet lengths with tcpdump.

So remove the extra ETH_HLEN from the veth byte count totals.

Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet ipv6: Fix duplicate /proc/sys/net/ipv6/neigh directory entries.
Eric W. Biederman [Tue, 22 Mar 2011 01:23:34 +0000 (18:23 -0700)]
net ipv6: Fix duplicate /proc/sys/net/ipv6/neigh directory entries.

When I was fixing issues with unregisgtering tables under /proc/sys/net/ipv6/neigh
by adding a mount point it appears I missed a critical ordering issue, in the
ipv6 initialization.  I had not realized that ipv6_sysctl_register is called
at the very end of the ipv6 initialization and in particular after we call
neigh_sysctl_register from ndisc_init.

"neigh" needs to be initialized in ipv6_static_sysctl_register which is
the first ipv6 table to initialized, and definitely before ndisc_init.
This removes the weirdness of duplicate tables while still providing a
"neigh" mount point which prevents races in sysctl unregistering.

This was initially reported at https://bugzilla.kernel.org/show_bug.cgi?id=31232
Reported-by: sunkan@zappa.cx
Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomacvlan: Fix use after free of struct macvlan_port.
Eric W. Biederman [Tue, 22 Mar 2011 01:22:22 +0000 (18:22 -0700)]
macvlan: Fix use after free of struct macvlan_port.

When the macvlan driver was extended to call unregisgter_netdevice_queue
in 23289a37e2b127dfc4de1313fba15bb4c9f0cd5b, a use after free of struct
macvlan_port was introduced.  The code in dellink relied on unregister_netdevice
actually unregistering the net device so it would be safe to free macvlan_port.

Since unregister_netdevice_queue can just queue up the unregister instead of
performing the unregiser immediately we free the macvlan_port too soon and
then the code in macvlan_stop removes the macaddress for the set of macaddress
to listen for and uses memory that has already been freed.

To fix this add a reference count to track when it is safe to free the macvlan_port
and move the call of macvlan_port_destroy into macvlan_uninit which is guaranteed
to be called after the final macvlan_port_close.

Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet: fix incorrect spelling in drop monitor protocol
Neil Horman [Tue, 22 Mar 2011 01:20:26 +0000 (18:20 -0700)]
net: fix incorrect spelling in drop monitor protocol

It was pointed out to me recently that my spelling could be better :)

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agocan: c_can: Do basic c_can configuration _before_ enabling the interrupts
Jan Altenberg [Tue, 22 Mar 2011 01:19:26 +0000 (18:19 -0700)]
can: c_can: Do basic c_can configuration _before_ enabling the interrupts

I ran into some trouble while testing the SocketCAN driver for the BOSCH
C_CAN controller. The interface is not correctly initialized, if I put
some CAN traffic on the line, _while_ the interface is being started
(which means: the interface doesn't come up correcty, if there's some RX
traffic while doing 'ifconfig can0 up').

The current implementation enables the controller interrupts _before_
doing the basic c_can configuration. I think, this should be done the
other way round.

The patch below fixes things for me.

Signed-off-by: Jan Altenberg <jan@linutronix.de>
Acked-by: Kurt Van Dijck <kurt.van.dijck@eia.be>
Acked-by: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agonet/appletalk: fix atalk_release use after free
Arnd Bergmann [Tue, 22 Mar 2011 01:18:00 +0000 (18:18 -0700)]
net/appletalk: fix atalk_release use after free

The BKL removal in appletalk introduced a use-after-free problem,
where atalk_destroy_socket frees a sock, but we still release
the socket lock on it.

An easy fix is to take an extra reference on the sock and sock_put
it when returning from atalk_release.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoipx: fix ipx_release()
Eric Dumazet [Tue, 22 Mar 2011 01:16:39 +0000 (18:16 -0700)]
ipx: fix ipx_release()

Commit b0d0d915d1d1a0 (remove the BKL) added a regression, because
sock_put() can free memory while we are going to use it later.

Fix is to delay sock_put() _after_ release_sock().

Reported-by: Ingo Molnar <mingo@elte.hu>
Tested-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agosnmp: SNMP_UPD_PO_STATS_BH() always called from softirq
Eric Dumazet [Tue, 22 Mar 2011 01:12:54 +0000 (18:12 -0700)]
snmp: SNMP_UPD_PO_STATS_BH() always called from softirq

We dont need to test if we run from softirq context, we definitely are.

This saves few instructions in ip_rcv() & ip_rcv_finish()

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agol2tp: fix possible oops on l2tp_eth module unload
James Chapman [Tue, 22 Mar 2011 01:10:25 +0000 (18:10 -0700)]
l2tp: fix possible oops on l2tp_eth module unload

A struct used in the l2tp_eth driver for registering network namespace
ops was incorrectly marked as __net_initdata, leading to oops when
module unloaded.

BUG: unable to handle kernel paging request at ffffffffa00ec098
IP: [<ffffffff8123dbd8>] ops_exit_list+0x7/0x4b
PGD 142d067 PUD 1431063 PMD 195da8067 PTE 0
Oops: 0000 [#1] SMP
last sysfs file: /sys/module/l2tp_eth/refcnt
Call Trace:
 [<ffffffff8123dc94>] ? unregister_pernet_operations+0x32/0x93
 [<ffffffff8123dd20>] ? unregister_pernet_device+0x2b/0x38
 [<ffffffff81068b6e>] ? sys_delete_module+0x1b8/0x222
 [<ffffffff810c7300>] ? do_munmap+0x254/0x318
 [<ffffffff812c64e5>] ? page_fault+0x25/0x30
 [<ffffffff812c6952>] ? system_call_fastpath+0x16/0x1b

Signed-off-by: James Chapman <jchapman@katalix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoxfrm: Fix initialize repl field of struct xfrm_state
Wei Yongjun [Tue, 22 Mar 2011 01:08:28 +0000 (18:08 -0700)]
xfrm: Fix initialize repl field of struct xfrm_state

Commit 'xfrm: Move IPsec replay detection functions to a separate file'
  (9fdc4883d92d20842c5acea77a4a21bb1574b495)
introduce repl field to struct xfrm_state, and only initialize it
under SA's netlink create path, the other path, such as pf_key,
ipcomp/ipcomp6 etc, the repl field remaining uninitialize. So if
the SA is created by pf_key, any input packet with SA's encryption
algorithm will cause panic.

    int xfrm_input()
    {
        ...
        x->repl->advance(x, seq);
        ...
    }

This patch fixed it by introduce new function __xfrm_init_state().

Pid: 0, comm: swapper Not tainted 2.6.38-next+ #14 Bochs Bochs
EIP: 0060:[<c078e5d5>] EFLAGS: 00010206 CPU: 0
EIP is at xfrm_input+0x31c/0x4cc
EAX: dd839c00 EBX: 00000084 ECX: 00000000 EDX: 01000000
ESI: dd839c00 EDI: de3a0780 EBP: dec1de88 ESP: dec1de64
 DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068
Process swapper (pid: 0, ti=dec1c000 task=c09c0f20 task.ti=c0992000)
Stack:
 00000000 00000000 00000002 c0ba27c0 00100000 01000000 de3a0798 c0ba27c0
 00000033 dec1de98 c0786848 00000000 de3a0780 dec1dea4 c0786868 00000000
 dec1debc c074ee56 e1da6b8c de3a0780 c074ed44 de3a07a8 dec1decc c074ef32
Call Trace:
 [<c0786848>] xfrm4_rcv_encap+0x22/0x27
 [<c0786868>] xfrm4_rcv+0x1b/0x1d
 [<c074ee56>] ip_local_deliver_finish+0x112/0x1b1
 [<c074ed44>] ? ip_local_deliver_finish+0x0/0x1b1
 [<c074ef32>] NF_HOOK.clone.1+0x3d/0x44
 [<c074ef77>] ip_local_deliver+0x3e/0x44
 [<c074ed44>] ? ip_local_deliver_finish+0x0/0x1b1
 [<c074ec03>] ip_rcv_finish+0x30a/0x332
 [<c074e8f9>] ? ip_rcv_finish+0x0/0x332
 [<c074ef32>] NF_HOOK.clone.1+0x3d/0x44
 [<c074f188>] ip_rcv+0x20b/0x247
 [<c074e8f9>] ? ip_rcv_finish+0x0/0x332
 [<c072797d>] __netif_receive_skb+0x373/0x399
 [<c0727bc1>] netif_receive_skb+0x4b/0x51
 [<e0817e2a>] cp_rx_poll+0x210/0x2c4 [8139cp]
 [<c072818f>] net_rx_action+0x9a/0x17d
 [<c0445b5c>] __do_softirq+0xa1/0x149
 [<c0445abb>] ? __do_softirq+0x0/0x149

Signed-off-by: Wei Yongjun <yjwei@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agomac80211: initialize sta->last_rx in sta_info_alloc
Felix Fietkau [Mon, 21 Mar 2011 19:01:00 +0000 (20:01 +0100)]
mac80211: initialize sta->last_rx in sta_info_alloc

This field is used to determine the inactivity time. When in AP mode,
hostapd uses it for kicking out inactive clients after a while. Without this
patch, hostapd immediately deauthenticates a new client if it checks the
inactivity time before the client sends its first data frame.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoath9k: Fix kernel panic in AR2427
Mohammed Shafi Shajakhan [Mon, 21 Mar 2011 12:57:21 +0000 (18:27 +0530)]
ath9k: Fix kernel panic in AR2427

Kernel panic occurs just after AR2427 establishes connection with AP.
Unless aggregation is enabled we don't initialize the TID structure.
Thus accesing the elements of the TID structure when aggregation is
disabled, leads to NULL pointer dereferencing.

[  191.320358] Call Trace:
[  191.320364]  [<fd250ea7>] ? ath9k_tx+0xa7/0x200 [ath9k]
[  191.320376]  [<fd1ec7fc>] ? __ieee80211_tx+0x5c/0x1e0 [mac80211]
[  191.320386]  [<fd1edd2b>] ? ieee80211_tx+0x7b/0x90 [mac80211]
[  191.320395]  [<fd1edddd>] ? ieee80211_xmit+0x9d/0x1d0 [mac80211]
[  191.320401]  [<c014218f>] ? wake_up_state+0xf/0x20
[  191.320405]  [<c015dbc8>] ? signal_wake_up+0x28/0x40
[  191.320410]  [<c012a578>] ? default_spin_lock_flags+0x8/0x10
[  191.320420]  [<fd1ee308>] ? ieee80211_subif_start_xmit+0x2e8/0x7c0
[mac80211]
[  191.320425]  [<c058f905>] ? do_page_fault+0x295/0x3a0
[  191.320431]  [<c04c4a3d>] ? dev_hard_start_xmit+0x1ad/0x210
[  191.320436]  [<c04d96b5>] ? sch_direct_xmit+0x105/0x170
[  191.320445]  [<fd1f161a>] ? get_sta_flags+0x2a/0x40 [mac80211]
[  191.320449]  [<c04c780f>] ? dev_queue_xmit+0x37f/0x4b0
[  191.320452]  [<c04d75b0>] ? eth_header+0x0/0xb0
[  191.320456]  [<c04cc479>] ? neigh_resolve_output+0xe9/0x310
[  191.320461]  [<c053d295>] ? ip6_output_finish+0xa5/0x110
[  191.320464]  [<c053e354>] ? ip6_output2+0x134/0x250
[  191.320468]  [<c053f7dd>] ? ip6_output+0x6d/0x100
[  191.320471]  [<c0559665>] ? mld_sendpack+0x395/0x3e0
[  191.320475]  [<c0557f81>] ? add_grhead+0x31/0xa0
[  191.320478]  [<c055a83c>] ? mld_send_cr+0x1bc/0x2b0
[  191.320482]  [<c01535d9>] ? irq_exit+0x39/0x70
[  191.320485]  [<c055a940>] ? mld_ifc_timer_expire+0x10/0x40
[  191.320489]  [<c015b92e>] ? run_timer_softirq+0x13e/0x2c0
[  191.320493]  [<c0103a30>] ? common_interrupt+0x30/0x40
[  191.320498]  [<c055a930>] ? mld_ifc_timer_expire+0x0/0x40
[  191.320502]  [<c0153358>] ? __do_softirq+0x98/0x1b0
[  191.320506]  [<c01534b5>] ? do_softirq+0x45/0x50
[  191.320509]  [<c0153605>] ? irq_exit+0x65/0x70
[  191.320513]  [<c05917dc>] ? smp_apic_timer_interrupt+0x5c/0x8b
[  191.320516]  [<c0103df1>] ? apic_timer_interrupt+0x31/0x40
[  191.320521]  [<c016007b>] ? k_getrusage+0x12b/0x2f0
[  191.320525]  [<c039e384>] ? acpi_idle_enter_simple+0x117/0x148
[  191.320529]  [<c04a20da>] ? cpuidle_idle_call+0x7a/0x100
[  191.320532]  [<c01021d4>] ? cpu_idle+0x94/0xd0
[  191.320536]  [<c057ab88>] ? rest_init+0x58/0x60
[  191.320541]  [<c07a58ec>] ? start_kernel+0x351/0x357
[  191.320544]  [<c07a53c7>] ? unknown_bootoption+0x0/0x19e
[  191.320548]  [<c07a50aa>] ? i386_start_kernel+0xaa/0xb1
[  191.320550] Code: 03 66 3d 00 03 0f 84 7c 02 00 00 83 c3 18 0f b6 03
8b 4d e0 89 c3 83 e3 0f 6b c3 48 89 5d d8 8d 04 06 8d 50 0c 89 55 d0 8b
40 20 <8b> 00 3b 01 0f 85 8e 02 00 00 f6 47 20 40 0f 84 29 ff ff ff 8b
[  191.320634] EIP: [<fd2586d4>] ath_tx_start+0x474/0x770 [ath9k] SS:ESP
0068:c0761a90
[  191.320642] CR2: 0000000000000000
[  191.320647] ---[ end trace 9296ef23b9076ece ]---
[  191.320650] Kernel panic - not syncing: Fatal exception in interrupt

Cc: stable@kernel.org
Signed-off-by: Mohammed Shafi Shajakhan <mshajakhan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agowireless: fix 80211 kernel-doc warnings
Randy Dunlap [Fri, 18 Mar 2011 16:33:02 +0000 (09:33 -0700)]
wireless: fix 80211 kernel-doc warnings

Fix many of each of these warnings:

Warning(include/net/cfg80211.h:519): No description found for parameter 'rxrate'
Warning(include/net/mac80211.h:1163): bad line:

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agortlwifi: Fix memset argument order
Ilia Mirkin [Thu, 17 Mar 2011 18:08:57 +0000 (14:08 -0400)]
rtlwifi: Fix memset argument order

These were found using the following semantic match:
// <smpl>
@@expression E1; type T;@@
* memset(E1, ... * sizeof(T) * ..., ...);
// </smpl>

Also take this opportunity to remove the unnecessary void* casts.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Tested-by: Larry Finger <Larry.Finger@lwfinger.net>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Add unknown Toshiba device
Peter Lemenkov [Wed, 16 Mar 2011 14:12:18 +0000 (17:12 +0300)]
rt2x00: Add unknown Toshiba device

Add unknown Toshiba device, mentioned in rt3572sta sources, under
CONFIG_RT2800USB_UNKNOWN

Signed-off-by: Peter Lemenkov <lemenkov@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Add 2L Central Europe BV 8070
Peter Lemenkov [Wed, 16 Mar 2011 14:12:17 +0000 (17:12 +0300)]
rt2x00: Add 2L Central Europe BV 8070

Add 2L Central Europe BV 8070 under CONFIG_RT2800USB_UNKNOWN

Signed-off-by: Peter Lemenkov <lemenkov@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Add Planex Communications, Inc. RT8070
Peter Lemenkov [Wed, 16 Mar 2011 14:12:16 +0000 (17:12 +0300)]
rt2x00: Add Planex Communications, Inc. RT8070

Add Planex Communications, Inc. RT8070 under CONFIG_RT2800USB_UNKNOWN

Signed-off-by: Peter Lemenkov <lemenkov@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agort2x00: Add rt2870 device id
Peter Lemenkov [Wed, 16 Mar 2011 14:12:15 +0000 (17:12 +0300)]
rt2x00: Add rt2870 device id

Add ID for Asus USB-N11 Wi-FI adapter. Tested by me.

Signed-off-by: Peter Lemenkov <lemenkov@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoorinoco: Maintain lock until entry removed from list
Joe Gunn [Tue, 15 Mar 2011 15:41:19 +0000 (08:41 -0700)]
orinoco: Maintain lock until entry removed from list

Removing an entry from the scan_list should be performed while holding
the lock.

Signed-off-by: Joseph J. Gunn <armadefuego@yahoo.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoiwlwifi: missing unlock on error path
Dan Carpenter [Tue, 15 Mar 2011 07:01:58 +0000 (10:01 +0300)]
iwlwifi: missing unlock on error path

We should unlock here instead of returning -EINVAL directly.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agozd1211rw: TrendNet TEW-509UB id added
Tõnu Samuel [Sat, 12 Mar 2011 09:29:25 +0000 (11:29 +0200)]
zd1211rw: TrendNet TEW-509UB id added

Signed-off-by: Tõnu Samuel <tonu@jes.ee>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
13 years agoMerge branch 'vhost-net-next' of git://git.kernel.org/pub/scm/linux/kernel/git/mst...
David S. Miller [Sun, 20 Mar 2011 21:35:09 +0000 (14:35 -0700)]
Merge branch 'vhost-net-next' of git://git./linux/kernel/git/mst/vhost

13 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/kaber/nf-2.6
David S. Miller [Sun, 20 Mar 2011 20:42:25 +0000 (13:42 -0700)]
Merge branch 'master' of git://git./linux/kernel/git/kaber/nf-2.6

13 years agonetfilter: ipt_CLUSTERIP: fix buffer overflow
Vasiliy Kulikov [Sun, 20 Mar 2011 14:42:52 +0000 (15:42 +0100)]
netfilter: ipt_CLUSTERIP: fix buffer overflow

'buffer' string is copied from userspace.  It is not checked whether it is
zero terminated.  This may lead to overflow inside of simple_strtoul().
Changli Gao suggested to copy not more than user supplied 'size' bytes.

It was introduced before the git epoch.  Files "ipt_CLUSTERIP/*" are
root writable only by default, however, on some setups permissions might be
relaxed to e.g. network admin user.

Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Acked-by: Changli Gao <xiaosuo@gmail.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
13 years agonetfilter: xtables: fix reentrancy
Eric Dumazet [Sun, 20 Mar 2011 14:40:06 +0000 (15:40 +0100)]
netfilter: xtables: fix reentrancy

commit f3c5c1bfd4308 (make ip_tables reentrant) introduced a race in
handling the stackptr restore, at the end of ipt_do_table()

We should do it before the call to xt_info_rdunlock_bh(), or we allow
cpu preemption and another cpu overwrites stackptr of original one.

A second fix is to change the underflow test to check the origptr value
instead of 0 to detect underflow, or else we allow a jump from different
hooks.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>
13 years agonetfilter: ipset: fix checking the type revision at create command
Jozsef Kadlecsik [Sun, 20 Mar 2011 14:35:01 +0000 (15:35 +0100)]
netfilter: ipset: fix checking the type revision at create command

The revision of the set type was not checked at the create command: if the
userspace sent a valid set type but with not supported revision number,
it'd create a loop.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
13 years agonetfilter: ipset: fix address ranges at hash:*port* types
Jozsef Kadlecsik [Sun, 20 Mar 2011 14:33:26 +0000 (15:33 +0100)]
netfilter: ipset: fix address ranges at hash:*port* types

The hash:*port* types with IPv4 silently ignored when address ranges
with non TCP/UDP were added/deleted from the set and used the first
address from the range only.

Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Patrick McHardy <kaber@trash.net>
13 years agoniu: Rename NIU parent platform device name to fix conflict.
David S. Miller [Sun, 20 Mar 2011 06:06:33 +0000 (23:06 -0700)]
niu: Rename NIU parent platform device name to fix conflict.

When the OF device driver bits were converted over to the platform
device infrastructure in commit 74888760d40b3ac9054f9c5fa07b566c0676ba2d
("dt/net: Eliminate users of of_platform_{,un}register_driver") we
inadvertantly created probing problems in the OF case.

The NIU driver creates a dummy platform device to represent the
board that contains one or more child NIU devices.  Unfortunately
we use the same name, "niu", as the OF device driver itself uses.

The result is that we try to probe the dummy "niu" parent device we
create, and since it has a NULL ofdevice pointer etc. everything
explodes:

[783019.128243] niu: niu.c:v1.1 (Apr 22, 2010)
[783019.128810] Unable to handle kernel NULL pointer dereference
[783019.128949] tsk->{mm,active_mm}->context = 000000000000039e
[783019.129078] tsk->{mm,active_mm}->pgd = fffff803afc5a000
[783019.129206]               \|/ ____ \|/
[783019.129213]               "@'/ .. \`@"
[783019.129220]               /_| \__/ |_\
[783019.129226]                  \__U_/
[783019.129378] modprobe(2004): Oops [#1]
[783019.129423] TSTATE: 0000000011001602 TPC: 0000000010052ff8 TNPC: 000000000061bbb4 Y: 00000000    Not tainted
[783019.129542] TPC: <niu_of_probe+0x3c/0x2dc [niu]>
[783019.129624] g0: 8080000000000000 g1: 0000000000000000 g2: 0000000010056000 g3: 0000000000000002
[783019.129733] g4: fffff803fc1da0c0 g5: fffff800441e2000 g6: fffff803fba84000 g7: 0000000000000000
[783019.129842] o0: fffff803fe7df010 o1: 0000000010055700 o2: 0000000000000000 o3: fffff803fbacaca0
[783019.129951] o4: 0000000000000080 o5: 0000000000777908 sp: fffff803fba866e1 ret_pc: 0000000010052ff4
[783019.130083] RPC: <niu_of_probe+0x38/0x2dc [niu]>
[783019.130165] l0: fffff803fe7df010 l1: fffff803fbacafc0 l2: fffff803fbacaca0 l3: ffffffffffffffed
[783019.130273] l4: 0000000000000000 l5: 000000007fffffff l6: fffff803fba86f40 l7: 0000000000000001
[783019.130382] i0: fffff803fe7df000 i1: fffff803fc20aba0 i2: 0000000000000000 i3: 0000000000000001
[783019.130490] i4: 0000000000000000 i5: 0000000000000000 i6: fffff803fba867a1 i7: 000000000062038c
[783019.130614] I7: <platform_drv_probe+0xc/0x20>

Fix by simply renaming the parent device to "niu-board".

Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agor8169: fix a bug in rtl8169_init_phy()
Eric Dumazet [Sat, 19 Mar 2011 05:39:11 +0000 (05:39 +0000)]
r8169: fix a bug in rtl8169_init_phy()

commit 54405cde7624 (r8169: support control of advertising.)
introduced a bug in rtl8169_init_phy()

Reported-by: Piotr Hosowicz <piotr@hosowicz.com>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Oliver Neukum <oliver@neukum.org>
Cc: Francois Romieu <romieu@fr.zoreil.com>
Tested-by: Anca Emanuel <anca.emanuel@gmail.com>
Tested-by: Piotr Hosowicz <piotr@hosowicz.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agobonding: fix a typo in a comment
Nicolas de Pesloüan [Sat, 19 Mar 2011 20:36:18 +0000 (13:36 -0700)]
bonding: fix a typo in a comment

Signed-off-by: Nicolas de Pesloüan <nicolas.2p.debian@free.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoftmac100: use resource_size()
Dan Carpenter [Sat, 19 Mar 2011 04:53:03 +0000 (21:53 -0700)]
ftmac100: use resource_size()

The calculation is off-by-one.  It should be "end - start + 1".  This
patch fixes it to use resource_size() instead.  Oddly, the code already
uses resource size correctly a couple lines earlier when it calls
request_mem_region() for this memory.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoheaders: use __aligned_xx types for userspace
Mike Frysinger [Fri, 18 Mar 2011 08:50:37 +0000 (08:50 +0000)]
headers: use __aligned_xx types for userspace

Now that we finally have __aligned_xx exported to userspace, convert
the headers that get exported over to the proper type.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agobridge: Reset IPCB when entering IP stack on NF_FORWARD
Herbert Xu [Fri, 18 Mar 2011 05:27:28 +0000 (05:27 +0000)]
bridge: Reset IPCB when entering IP stack on NF_FORWARD

Whenever we enter the IP stack proper from bridge netfilter we
need to ensure that the skb is in a form the IP stack expects
it to be in.

The entry point on NF_FORWARD did not meet the requirements of
the IP stack, therefore leading to potential crashes/panics.

This patch fixes the problem.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agovlan: should take into account needed_headroom
Eric Dumazet [Fri, 18 Mar 2011 00:27:27 +0000 (00:27 +0000)]
vlan: should take into account needed_headroom

Commit c95b819ad7 (gre: Use needed_headroom)
made gre use needed_headroom instead of hard_header_len

This uncover a bug in vlan code.

We should make sure vlan devices take into account their
real_dev->needed_headroom or we risk a crash in ipgre_header(), because
we dont have enough room to push IP header in skb.

Reported-by: Diddi Oscarsson <diddi@diddi.se>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoethtool: Compat handling for struct ethtool_rxnfc
Ben Hutchings [Thu, 17 Mar 2011 07:34:32 +0000 (07:34 +0000)]
ethtool: Compat handling for struct ethtool_rxnfc

This structure was accidentally defined such that its layout can
differ between 32-bit and 64-bit processes.  Add compat structure
definitions and an ioctl wrapper function.

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Acked-by: Alexander Duyck <alexander.h.duyck@intel.com>
Cc: stable@kernel.org [2.6.30+]
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoethtool: __ethtool_set_sg: check for function pointer before using it
Roger Luethi [Thu, 17 Mar 2011 06:37:21 +0000 (06:37 +0000)]
ethtool: __ethtool_set_sg: check for function pointer before using it

__ethtool_set_sg does not check if dev->ethtool_ops->set_sg is defined
which can result in a NULL pointer dereference when ethtool is used to
change SG settings for drivers without SG support.

Signed-off-by: Roger Luethi <rl@hellgate.ch>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoeconet: 4 byte infoleak to the network
Vasiliy Kulikov [Thu, 17 Mar 2011 01:40:10 +0000 (01:40 +0000)]
econet: 4 byte infoleak to the network

struct aunhdr has 4 padding bytes between 'pad' and 'handle' fields on
x86_64.  These bytes are not initialized in the variable 'ah' before
sending 'ah' to the network.  This leads to 4 bytes kernel stack
infoleak.

This bug was introduced before the git epoch.

Signed-off-by: Vasiliy Kulikov <segoon@openwall.com>
Acked-by: Phil Blundell <philb@gnu.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agogianfar: Fall back to software tcp/udp checksum on older controllers
Alex Dubov [Wed, 16 Mar 2011 17:57:13 +0000 (17:57 +0000)]
gianfar: Fall back to software tcp/udp checksum on older controllers

As specified by errata eTSEC49 of MPC8548 and errata eTSEC12 of MPC83xx,
older revisions of gianfar controllers will be unable to calculate a TCP/UDP
packet checksum for some alignments of the appropriate FCB. This patch checks
for FCB alignment on such controllers and falls back to software checksumming
if the alignment is known to be bad.

Signed-off-by: Alex Dubov <oakad@yahoo.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
13 years agoMerge branch 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes...
Linus Torvalds [Fri, 18 Mar 2011 17:56:44 +0000 (10:56 -0700)]
Merge branch 'linux-next' of git://git./linux/kernel/git/jbarnes/pci-2.6

* 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci-2.6:
  PCI: label: remove #include of ACPI header to avoid warnings
  PCI: label: Fix compilation error when CONFIG_ACPI is unset
  PCI: pre-allocate additional resources to devices only after successful allocation of essential resources.
  PCI: introduce reset_resource()
  PCI: data structure agnostic free list function
  PCI: refactor io size calculation code
  PCI: do not create quirk I/O regions below PCIBIOS_MIN_IO for ICH
  PCI hotplug: acpiphp: set current_state to D0 in register_slot
  PCI: Export ACPI _DSM provided firmware instance number and string name to sysfs
  PCI: add more checking to ICH region quirks
  PCI: aer-inject: Override PCIe AER Mask Registers
  PCI: fix tlan build when CONFIG_PCI is not enabled
  PCI: remove quirk for pre-production systems
  PCI: Avoid potential NULL pointer dereference in pci_scan_bridge
  PCI/lpc: irq and pci_ids patch for Intel DH89xxCC DeviceIDs
  PCI: sysfs: Fix failure path for addition of "vpd" attribute

13 years agoMerge branch 'spi/next' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Fri, 18 Mar 2011 17:56:02 +0000 (10:56 -0700)]
Merge branch 'spi/next' of git://git.secretlab.ca/git/linux-2.6

* 'spi/next' of git://git.secretlab.ca/git/linux-2.6: (34 commits)
  spi/dw_spi: move dw_spi.h into drivers/spi
  spi/dw_spi: Fix missing header
  gpio/langwell: Clear edge bit before handling
  gpio/langwell: Simplify demux loop
  gpio/langwell: Convert irq name space
  gpio/langwell: Fix broken irq_eoi change.
  gpio; Make Intel chipset gpio drivers depend on x86
  gpio/cs5535-gpio: Fix section mismatch
  spi/rtc-{ds1390,ds3234,m41t94}: Use spi_get_drvdata() for SPI devices
  spi/davinci: Support DMA transfers larger than 65535 words
  spi/davinci: Use correct length parameter to dma_map_single calls
  gpio: Use __devexit at necessary places
  gpio: add MODULE_DEVICE_TABLE to pch_gpio and ml_ioh_gpio
  gpio/mcp23s08: support mcp23s17 variant
  of_mmc_spi: add card detect irq support
  spi/omap_mcspi: catch xfers of non-multiple SPI word size
  spi/omap_mcspi: Off-by-one error in finding the right divisor
  gpio/pca953x: Fix wrong pointer type
  spi/pl022: rid dangling labels
  spi: add support for SuperH SPI
  ...

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm
Linus Torvalds [Fri, 18 Mar 2011 17:55:11 +0000 (10:55 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/teigland/dlm

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/teigland/dlm:
  dlm: use alloc_workqueue function
  dlm: increase default hash table sizes
  dlm: record full callback state

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
Linus Torvalds [Fri, 18 Mar 2011 17:51:11 +0000 (10:51 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  fs: call security_d_instantiate in d_obtain_alias V2
  lose 'mounting_here' argument in ->d_manage()
  don't pass 'mounting_here' flag to follow_down()
  change the locking order for namespace_sem
  fix deadlock in pivot_root()
  vfs: split off vfsmount-related parts of vfs_kern_mount()
  Some fixes for pstore
  kill simple_set_mnt()

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bcopeland...
Linus Torvalds [Fri, 18 Mar 2011 17:50:52 +0000 (10:50 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/bcopeland/omfs

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/bcopeland/omfs:
  omfs: make readdir stop when filldir says so
  omfs: merge unlink() and rmdir(), close leak in rename()
  omfs: stop playing silly buggers with omfs_unlink() in ->rename()
  omfs: rename() needs to mark old_inode dirty after ctime update

13 years agoMerge branch 'linux-next' of git://git.infradead.org/ubifs-2.6
Linus Torvalds [Fri, 18 Mar 2011 17:50:27 +0000 (10:50 -0700)]
Merge branch 'linux-next' of git://git.infradead.org/ubifs-2.6

* 'linux-next' of git://git.infradead.org/ubifs-2.6: (25 commits)
  UBIFS: clean-up commentaries
  UBIFS: save 128KiB or more RAM
  UBIFS: allocate orphans scan buffer on demand
  UBIFS: allocate lpt dump buffer on demand
  UBIFS: allocate ltab checking buffer on demand
  UBIFS: allocate scanning buffer on demand
  UBIFS: allocate dump buffer on demand
  UBIFS: do not check data crc by default
  UBIFS: simplify UBIFS Kconfig menu
  UBIFS: print max. index node size
  UBIFS: handle allocation failures in UBIFS write path
  UBIFS: use max_write_size during recovery
  UBIFS: use max_write_size for write-buffers
  UBIFS: introduce write-buffer size field
  UBI: incorporate LEB offset information
  UBIFS: incorporate maximum write size
  UBI: provide LEB offset information
  UBI: incorporate maximum write size
  UBIFS: fix LEB number in printk
  UBIFS: restrict world-writable debugfs files
  ...

13 years agoMerge branch 'linux-next' of git://git.infradead.org/ubi-2.6
Linus Torvalds [Fri, 18 Mar 2011 17:50:02 +0000 (10:50 -0700)]
Merge branch 'linux-next' of git://git.infradead.org/ubi-2.6

* 'linux-next' of git://git.infradead.org/ubi-2.6:
  UBI: make tests modes dynamic
  UBI: make self-checks dynamic
  UBI: make debugging messages dynamic
  UBI: remove UBI_IO_DEBUG macro
  UBI: kill debugging buffer
  UBI: allocate erase checking buffer on demand
  UBI: allocate write checking buffer on demand
  UBI: always re-read in case of read failures
  UBI: cleanup comments about corrupted PEBs
  UBI: add slab cache for ubi_scan_leb objects
  UBI: use raw mtd read function in debugging code
  UBI: try to reveal buggy MTD drivers
  UBI: add a commentary about allocating VID header buffer on stack
  UBI: cleanup LEB start calculations
  UBI: fix NOR erase preparation quirk

13 years agoMAINTAINERS: Add serial
Alan Cox [Fri, 18 Mar 2011 13:56:14 +0000 (13:56 +0000)]
MAINTAINERS: Add serial

Add an entry for serial maintainers.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Fri, 18 Mar 2011 17:46:37 +0000 (10:46 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6: (308 commits)
  ALSA: sound/pci/asihpi: check adapter index in hpi_ioctl
  ALSA: aloop - Fix possible IRQ lock inversion
  ALSA: sound/core: merge list_del()/list_add_tail() to list_move_tail()
  ALSA: ctxfi - use list_move() instead of list_del()/list_add() combination
  ALSA: firewire - msleep needs delay.h
  ALSA: firewire-lib, firewire-speakers: handle packet queueing errors
  ALSA: firewire-lib: allocate DMA buffer separately
  ALSA: firewire-lib: use no-info SYT for packets without SYT sample
  ALSA: add LaCie FireWire Speakers/Griffin FireWave Surround driver
  ALSA: hda - Remove an unused variable in patch_realtek.c
  ALSA: hda - pin-adc-mux-dmic auto-configuration of 92HD8X codecs
  ALSA: hda - fix digital mic selection in mixer on 92HD8X codecs
  ALSA: hda - Move default input-src selection to init part
  ALSA: hda - Initialize special cases for input src in init phase
  ALSA: ctxfi - Clear input settings before initialization
  ALSA: ctxfi - Fix SPDIF status retrieval
  ALSA: ctxfi - Fix incorrect SPDIF status bit mask
  ALSA: ctxfi - Fix microphone boost codes/comments
  ALSA: atiixp - Fix wrong time-out checks during ac-link reset
  ALSA: intel8x0m: append 'm' to "r_intel8x0"
  ...

13 years agoMerge branch 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 18 Mar 2011 17:45:21 +0000 (10:45 -0700)]
Merge branch 'x86-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  x86: Flush TLB if PGD entry is changed in i386 PAE mode
  x86, dumpstack: Correct stack dump info when frame pointer is available
  x86: Clean up csum-copy_64.S a bit
  x86: Fix common misspellings
  x86: Fix misspelling and align params
  x86: Use PentiumPro-optimized partial_csum() on VIA C7

13 years agoMerge branches 'irq-fixes-for-linus' and 'sched-fixes-for-linus' of git://git.kernel...
Linus Torvalds [Fri, 18 Mar 2011 17:44:05 +0000 (10:44 -0700)]
Merge branches 'irq-fixes-for-linus' and 'sched-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'irq-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  genirq: Fix incorrect unlock in __setup_irq()
  cris: Use generic show_interrupts()
  genirq: show_interrupts: Check desc->name before printing it blindly
  cris: Use accessor functions to set IRQ_PER_CPU flag
  cris: Fix irq conversion fallout

* 'sched-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  sched, kernel-doc: Fix runqueue_is_locked() description

13 years agoMerge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 18 Mar 2011 17:38:34 +0000 (10:38 -0700)]
Merge branch 'perf-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (30 commits)
  trace, filters: Initialize the match variable in process_ops() properly
  trace, documentation: Fix branch profiling location in debugfs
  oprofile, s390: Cleanups
  oprofile, s390: Remove hwsampler_files.c and merge it into init.c
  perf: Fix tear-down of inherited group events
  perf: Reorder & optimize perf_event_context to remove alignment padding on 64 bit builds
  perf: Handle stopped state with tracepoints
  perf: Fix the software events state check
  perf, powerpc: Handle events that raise an exception without overflowing
  perf, x86: Use INTEL_*_CONSTRAINT() for all PEBS event constraints
  perf, x86: Clean up SandyBridge PEBS events
  perf lock: Fix sorting by wait_min
  perf tools: Version incorrect with some versions of grep
  perf evlist: New command to list the names of events present in a perf.data file
  perf script: Add support for H/W and S/W events
  perf script: Add support for dumping symbols
  perf script: Support custom field selection for output
  perf script: Move printing of 'common' data from print_event and rename
  perf tracing: Remove print_graph_cpu and print_graph_proc from trace-event-parse
  perf script: Change process_event prototype
  ...

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial
Linus Torvalds [Fri, 18 Mar 2011 17:37:40 +0000 (10:37 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/trivial

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial: (47 commits)
  doc: CONFIG_UNEVICTABLE_LRU doesn't exist anymore
  Update cpuset info & webiste for cgroups
  dcdbas: force SMI to happen when expected
  arch/arm/Kconfig: remove one to many l's in the word.
  asm-generic/user.h: Fix spelling in comment
  drm: fix printk typo 'sracth'
  Remove one to many n's in a word
  Documentation/filesystems/romfs.txt: fixing link to genromfs
  drivers:scsi Change printk typo initate -> initiate
  serial, pch uart: Remove duplicate inclusion of linux/pci.h header
  fs/eventpoll.c: fix spelling
  mm: Fix out-of-date comments which refers non-existent functions
  drm: Fix printk typo 'failled'
  coh901318.c: Change initate to initiate.
  mbox-db5500.c Change initate to initiate.
  edac: correct i82975x error-info reported
  edac: correct i82975x mci initialisation
  edac: correct commented info
  fs: update comments to point correct document
  target: remove duplicate include of target/target_core_device.h from drivers/target/target_core_hba.c
  ...

Trivial conflict in fs/eventpoll.c (spelling vs addition)

13 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Fri, 18 Mar 2011 17:35:30 +0000 (10:35 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid: (48 commits)
  HID: add support for Logitech Driving Force Pro wheel
  HID: hid-ortek: remove spurious reference
  HID: add support for Ortek PKB-1700
  HID: roccat-koneplus: vorrect mode of sysfs attr 'sensor'
  HID: hid-ntrig: init settle and mode check
  HID: merge hid-egalax into hid-multitouch
  HID: hid-multitouch: Send events per slot if CONTACTCOUNT is missing
  HID: ntrig remove if and drop an indent
  HID: ACRUX - activate the device immediately after binding
  HID: ntrig: apply NO_INIT_REPORTS quirk
  HID: hid-magicmouse: Correct touch orientation direction
  HID: ntrig don't dereference unclaimed hidinput
  HID: Do not create input devices for feature reports
  HID: bt hidp: send Output reports using SET_REPORT on the Control channel
  HID: hid-sony.c: Fix sending Output reports to the Sixaxis
  HID: add support for Keytouch IEC 60945
  HID: Add HID Report Descriptor to sysfs
  HID: add IRTOUCH infrared USB to hid_have_special_driver
  HID: kernel oops in out_cleanup in function hidinput_connect
  HID: Add teletext/color keys - gyration remote - EU version (GYAR3101CKDE)
  ...

13 years agospi/dw_spi: move dw_spi.h into drivers/spi
Grant Likely [Mon, 28 Feb 2011 19:47:12 +0000 (12:47 -0700)]
spi/dw_spi: move dw_spi.h into drivers/spi

include/linux/dw_spi.h only includes driver internal data.  It doesn't
expose a platform_data configuration structure or similar (at least
nothing in-tree).  This patch moves the header into drivers/spi so
that the scope is limited to only the dw_spi_*.c driver files

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Cc: Feng Tang <feng.tang@intel.com>
Cc: spi-devel-general@lists.sourceforge.net
13 years agospi/dw_spi: Fix missing header
Jiri Slaby [Fri, 18 Mar 2011 09:41:17 +0000 (10:41 +0100)]
spi/dw_spi: Fix missing header

Currently, build on PPC dies with:
In file included from drivers/spi/dw_spi_mmio.c:16:
include/linux/spi/dw_spi.h:147: error: field ‘tx_sgl’ has incomplete type
include/linux/spi/dw_spi.h:149: error: field ‘rx_sgl’ has incomplete type

Add linux/scatterlist.h include to dw_spi.h, because we need to know
the contents of the structure.

Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
13 years agofs: call security_d_instantiate in d_obtain_alias V2
Josef Bacik [Fri, 19 Nov 2010 01:52:55 +0000 (20:52 -0500)]
fs: call security_d_instantiate in d_obtain_alias V2

While trying to track down some NFS problems with BTRFS, I kept noticing I was
getting -EACCESS for no apparent reason.  Eric Paris and printk() helped me
figure out that it was SELinux that was giving me grief, with the following
denial

type=AVC msg=audit(1290013638.413:95): avc:  denied  { 0x800000 } for  pid=1772
comm="nfsd" name="" dev=sda1 ino=256 scontext=system_u:system_r:kernel_t:s0
tcontext=system_u:object_r:unlabeled_t:s0 tclass=file

Turns out this is because in d_obtain_alias if we can't find an alias we create
one and do all the normal instantiation stuff, but we don't do the
security_d_instantiate.

Usually we are protected from getting a hashed dentry that hasn't yet run
security_d_instantiate() by the parent's i_mutex, but obviously this isn't an
option there, so in order to deal with the case that a second thread comes in
and finds our new dentry before we get to run security_d_instantiate(), we go
ahead and call it if we find a dentry already.  Eric assures me that this is ok
as the code checks to see if the dentry has been initialized already so calling
security_d_instantiate() against the same dentry multiple times is ok.  With
this patch I'm no longer getting errant -EACCESS values.

Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agolose 'mounting_here' argument in ->d_manage()
Al Viro [Fri, 18 Mar 2011 13:09:02 +0000 (09:09 -0400)]
lose 'mounting_here' argument in ->d_manage()

it's always false...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agotrace, filters: Initialize the match variable in process_ops() properly
Ingo Molnar [Fri, 18 Mar 2011 13:41:27 +0000 (14:41 +0100)]
trace, filters: Initialize the match variable in process_ops() properly

Make sure the 'match' variable always has a value.

Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Fri, 18 Mar 2011 13:31:43 +0000 (06:31 -0700)]
Merge branch 'next' of git://git./linux/kernel/git/benh/powerpc

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (62 commits)
  powerpc/85xx: Fix signedness bug in cache-sram
  powerpc/fsl: 85xx: document cache sram bindings
  powerpc/fsl: define binding for fsl mpic interrupt controllers
  powerpc/fsl_msi: Handle msi-available-ranges better
  drivers/serial/ucc_uart.c: Add of_node_put to avoid memory leak
  powerpc/85xx: Fix SPE float to integer conversion failure
  powerpc/85xx: Update sata controller compatible for p1022ds board
  ATA: Add FSL sata v2 controller support
  powerpc/mpc8xxx_gpio: simplify searching for 'fsl, qoriq-gpio' compatiable
  powerpc/8xx: remove obsolete mgsuvd board
  powerpc/82xx: rename and update mgcoge board support
  powerpc/83xx: rename and update kmeter1
  powerpc/85xx: Workaroudn e500 CPU erratum A005
  powerpc/fsl_pci: Add support for FSL PCIe controllers v2.x
  powerpc/85xx: Fix writing to spin table 'cpu-release-addr' on ppc64e
  powerpc/pseries: Disable MSI using new interface if possible
  powerpc: Enable GENERIC_HARDIRQS_NO_DEPRECATED.
  powerpc: core irq_data conversion.
  powerpc: sysdev/xilinx_intc irq_data conversion.
  powerpc: sysdev/uic irq_data conversion.
  ...

Fix up conflicts in arch/powerpc/sysdev/fsl_msi.c (due to getting rid of
of_platform_driver in arch/powerpc)

13 years agodon't pass 'mounting_here' flag to follow_down()
Al Viro [Fri, 18 Mar 2011 13:04:20 +0000 (09:04 -0400)]
don't pass 'mounting_here' flag to follow_down()

it's always false now

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agochange the locking order for namespace_sem
Al Viro [Fri, 18 Mar 2011 12:55:38 +0000 (08:55 -0400)]
change the locking order for namespace_sem

Have it nested inside ->i_mutex.  Instead of using follow_down()
under namespace_sem, followed by grabbing i_mutex and checking that
mountpoint to be is not dead, do the following:
grab i_mutex
check that it's not dead
grab namespace_sem
see if anything is mounted there
if not, we've won
otherwise
drop locks
put_path on what we had
replace with what's mounted
retry everything with new mountpoint to be

New helper (lock_mount()) does that.  do_add_mount(), do_move_mount(),
do_loopback() and pivot_root() switched to it; in case of the last
two that eliminates a race we used to have - original code didn't
do follow_down().

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agofix deadlock in pivot_root()
Al Viro [Fri, 18 Mar 2011 12:29:36 +0000 (08:29 -0400)]
fix deadlock in pivot_root()

Don't hold vfsmount_lock over the loop traversing ->mnt_parent;
do check_mnt(new.mnt) under namespace_sem instead; combined with
namespace_sem held over all that code it'll guarantee the stability
of ->mnt_parent chain all the way to the root.

Doing check_mnt() outside of namespace_sem in case of pivot_root()
is wrong anyway.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
13 years agox86: Flush TLB if PGD entry is changed in i386 PAE mode
Shaohua Li [Wed, 16 Mar 2011 03:37:29 +0000 (11:37 +0800)]
x86: Flush TLB if PGD entry is changed in i386 PAE mode

According to intel CPU manual, every time PGD entry is changed in i386 PAE
mode, we need do a full TLB flush. Current code follows this and there is
comment for this too in the code.

But current code misses the multi-threaded case. A changed page table
might be used by several CPUs, every such CPU should flush TLB. Usually
this isn't a problem, because we prepopulate all PGD entries at process
fork. But when the process does munmap and follows new mmap, this issue
will be triggered.

When it happens, some CPUs keep doing page faults:

  http://marc.info/?l=linux-kernel&m=129915020508238&w=2

Reported-by: Yasunori Goto<y-goto@jp.fujitsu.com>
Tested-by: Yasunori Goto<y-goto@jp.fujitsu.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Shaohua Li<shaohua.li@intel.com>
Cc: Mallick Asit K <asit.k.mallick@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mm <linux-mm@kvack.org>
Cc: stable <stable@kernel.org>
LKML-Reference: <1300246649.2337.95.camel@sli10-conroe>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agox86, dumpstack: Correct stack dump info when frame pointer is available
Namhyung Kim [Fri, 18 Mar 2011 02:40:06 +0000 (11:40 +0900)]
x86, dumpstack: Correct stack dump info when frame pointer is available

Current stack dump code scans entire stack and check each entry
contains a pointer to kernel code. If CONFIG_FRAME_POINTER=y it
could mark whether the pointer is valid or not based on value of
the frame pointer. Invalid entries could be preceded by '?' sign.

However this was not going to happen because scan start point
was always higher than the frame pointer so that they could not
meet.

Commit 9c0729dc8062 ("x86: Eliminate bp argument from the stack
tracing routines") delayed bp acquisition point, so the bp was
read in lower frame, thus all of the entries were marked
invalid.

This patch fixes this by reverting above commit while retaining
stack_frame() helper as suggested by Frederic Weisbecker.

End result looks like below:

before:

 [    3.508329] Call Trace:
 [    3.508551]  [<ffffffff814f35c9>] ? panic+0x91/0x199
 [    3.508662]  [<ffffffff814f3739>] ? printk+0x68/0x6a
 [    3.508770]  [<ffffffff81a981b2>] ? mount_block_root+0x257/0x26e
 [    3.508876]  [<ffffffff81a9821f>] ? mount_root+0x56/0x5a
 [    3.508975]  [<ffffffff81a98393>] ? prepare_namespace+0x170/0x1a9
 [    3.509216]  [<ffffffff81a9772b>] ? kernel_init+0x1d2/0x1e2
 [    3.509335]  [<ffffffff81003894>] ? kernel_thread_helper+0x4/0x10
 [    3.509442]  [<ffffffff814f6880>] ? restore_args+0x0/0x30
 [    3.509542]  [<ffffffff81a97559>] ? kernel_init+0x0/0x1e2
 [    3.509641]  [<ffffffff81003890>] ? kernel_thread_helper+0x0/0x10

after:

 [    3.522991] Call Trace:
 [    3.523351]  [<ffffffff814f35b9>] panic+0x91/0x199
 [    3.523468]  [<ffffffff814f3729>] ? printk+0x68/0x6a
 [    3.523576]  [<ffffffff81a981b2>] mount_block_root+0x257/0x26e
 [    3.523681]  [<ffffffff81a9821f>] mount_root+0x56/0x5a
 [    3.523780]  [<ffffffff81a98393>] prepare_namespace+0x170/0x1a9
 [    3.523885]  [<ffffffff81a9772b>] kernel_init+0x1d2/0x1e2
 [    3.523987]  [<ffffffff81003894>] kernel_thread_helper+0x4/0x10
 [    3.524228]  [<ffffffff814f6880>] ? restore_args+0x0/0x30
 [    3.524345]  [<ffffffff81a97559>] ? kernel_init+0x0/0x1e2
 [    3.524445]  [<ffffffff81003890>] ? kernel_thread_helper+0x0/0x10

 -v5:
   * fix build breakage with oprofile

 -v4:
   * use 0 instead of regs->bp
   * separate out printk changes

 -v3:
   * apply comment from Frederic
   * add a couple of printk fixes

Signed-off-by: Namhyung Kim <namhyung@gmail.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Soren Sandmann <ssp@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Robert Richter <robert.richter@amd.com>
LKML-Reference: <1300416006-3163-1-git-send-email-namhyung@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agox86: Clean up csum-copy_64.S a bit
Ingo Molnar [Fri, 18 Mar 2011 09:42:11 +0000 (10:42 +0100)]
x86: Clean up csum-copy_64.S a bit

The many stray whitespaces and other uncleanlinesses made this code
almost unreadable to me - so fix those.

No changes to the code.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agox86: Fix common misspellings
Lucas De Marchi [Thu, 17 Mar 2011 19:24:16 +0000 (16:24 -0300)]
x86: Fix common misspellings

They were generated by 'codespell' and then manually reviewed.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: trivial@kernel.org
LKML-Reference: <1300389856-1099-3-git-send-email-lucas.demarchi@profusion.mobi>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agox86: Fix misspelling and align params
Lucas De Marchi [Thu, 17 Mar 2011 19:24:15 +0000 (16:24 -0300)]
x86: Fix misspelling and align params

Fix 'upto' misspelling and align parameters.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Cc: trivial@kernel.org
LKML-Reference: <1300389856-1099-2-git-send-email-lucas.demarchi@profusion.mobi>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agoMerge branch 'linus' into x86/urgent
Ingo Molnar [Fri, 18 Mar 2011 09:38:53 +0000 (10:38 +0100)]
Merge branch 'linus' into x86/urgent

Merge reason: Merge upstream commits to avoid conflicts in upcoming patches.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
13 years agoMerge branch 'topic/misc' into for-linus
Takashi Iwai [Fri, 18 Mar 2011 06:39:08 +0000 (07:39 +0100)]
Merge branch 'topic/misc' into for-linus

13 years agoALSA: sound/pci/asihpi: check adapter index in hpi_ioctl
Dan Rosenberg [Thu, 17 Mar 2011 22:32:24 +0000 (18:32 -0400)]
ALSA: sound/pci/asihpi: check adapter index in hpi_ioctl

The user-supplied index into the adapters array needs to be checked, or
an out-of-bounds kernel pointer could be accessed and used, leading to
potentially exploitable memory corruption.

Signed-off-by: Dan Rosenberg <drosenberg@vsecurity.com>
Cc: <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
13 years agoALSA: aloop - Fix possible IRQ lock inversion
Takashi Iwai [Fri, 18 Mar 2011 06:31:53 +0000 (07:31 +0100)]
ALSA: aloop - Fix possible IRQ lock inversion

loopback_pos_update() can be called in the timer callback, thus the lock
held should be irq-safe.  Otherwise you'll get AB/BA deadlock together
with substream->self_group.lock.

Reported-and-tested-by: Knut Petersen <Knut_Petersen@t-online.de>
Cc: <stable@kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
13 years agogpio/langwell: Clear edge bit before handling
Thomas Gleixner [Thu, 17 Mar 2011 19:32:58 +0000 (19:32 +0000)]
gpio/langwell: Clear edge bit before handling

I don't have the specs for this beast, but it looks a lot like the PXA
GPIO block. Though I bet it's the same IP and the driver should have
reused the PXA code.

Acknowleding the edge detect status after handling one or more gpio
interrupts looks wrong. We might lose an edge which came in while we
handled the previous one.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Alek Du <alek.du@intel.com>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
13 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile
Linus Torvalds [Fri, 18 Mar 2011 02:34:12 +0000 (19:34 -0700)]
Merge git://git./linux/kernel/git/cmetcalf/linux-tile

* git://git.kernel.org/pub/scm/linux/kernel/git/cmetcalf/linux-tile: (27 commits)
  arch/tile: support newer binutils assembler shift semantics
  arch/tile: fix deadlock bugs in rwlock implementation
  drivers/edac: provide support for tile architecture
  tile on-chip network driver: sync up with latest fixes
  arch/tile: support 4KB page size as well as 64KB
  arch/tile: add some more VMSPLIT options and use consistent naming
  arch/tile: fix some comments and whitespace
  arch/tile: export some additional module symbols
  arch/tile: enhance existing finv_buffer_remote() routine
  arch/tile: fix two bugs in the backtracer code
  arch/tile: use extended assembly to inline __mb_incoherent()
  arch/tile: use a cleaner technique to enable interrupt for cpu_idle()
  arch/tile: sync up with <arch/sim.h> and <arch/sim_def.h> changes
  arch/tile: fix reversed test of strict_strtol() return value
  arch/tile: avoid a simulator warning during bootup
  arch/tile: export <asm/hardwall.h> to userspace
  arch/tile: warn and retry if an IPI is not accepted by the target cpu
  arch/tile: stop disabling INTCTRL_1 interrupts during hypervisor downcalls
  arch/tile: fix __ndelay etc to work better
  arch/tile: bug fix: exec'ed task thought it was still single-stepping
  ...

Fix up trivial conflict in arch/tile/kernel/vmlinux.lds.S (percpu
alignment vs section naming convention fix)